SoftBank Pepper Price 2026: Cost, Discontinued & Can You Buy One?

Is Pepper discontinued, and what would one cost you?

Classic Pepper is out of production and has been since 2020, but the story didn't end there: in February 2026 SoftBank's Japanese arm relaunched it as the rental-only Pepper+, and the company that inherited Pepper's European maker now builds only its sibling robot. Meanwhile the official US SoftBank Robotics site still presents Pepper as a current product, complete with consultation forms, which is exactly why so many pages tell you Pepper is alive and buyable. Here is the honest 2026 picture, checked against primary sources on July 22, 2026.

What actually happened to Pepper

Pepper's collapse is usually told as one event; it was a decade-long slide with precise dates.

Why did it die? Not because it was "creepy." Reuters' reporting cites weak demand, limited functionality, and friction between the French engineering unit and Tokyo; and restarting the stopped line would have been costly. The pause was real, the failure narrative came later, and the two comebacks below are why neither "dead" nor "back" is quite right.

The two comebacks, kept separate

Almost every 2025–2026 story about Pepper's afterlife conflates two unrelated threads.

Thread one: Pepper+ in Japan. SoftBank Robotics Group, the solvent Japanese company, launched Pepper+ on February 2, 2026: the first design and tablet refresh since 2014, an AI-agent conversation stack aimed at retail, reception and nursing care, and rental-only pricing from ¥79,800 a month (with an event plan at ¥150,000 and a care plan from ¥39,800). It is available in Japan only. The same announcement carried the program's nicest vindication: Guinness recognition of Pepper as the world's first mass-produced humanoid service robot.

Thread two: the Maxvision purchase. Maxvision, a listed Shenzhen border-security and AI firm, acquired Aldebaran's core technological assets and the Pepper and NAO intellectual property. Its French subsidiary Maxtronics is actively reviving NAO, Pepper's 58 cm desktop sibling: NAO6 is on sale, an NAO7 is in development, and aldebaran.com now redirects to maxtronics.com. Pepper, by contrast, appears on Maxtronics' site only as a legacy support archive. As of July 22, 2026 there is no announced plan to build Pepper again. So when a headline says "Pepper is saved," check which thread it means; the answer is usually NAO.

The $1,800 myth, decoded from the original price sheet

Pepper's launch price is one of tech's most misleading famous numbers. The ¥198,000 (~$1,800) sticker was roughly one-fifth of the advertised bundle, spelled out in SoftBank's own 2015 press release: a ¥14,800/month "Basic Plan" for the cloud and apps and a ¥9,800/month insurance pack, each running 36 months, for a three-year total of about ¥1.09 million (roughly $9,000 at 2015 rates). The same sheet notes standalone robot purchases could be applied for in shops, so "mandatory" overstates it slightly; in practice the Basic Plan was what made Pepper Pepper. The business version was rental-only at ¥55,000 a month on a non-cancellable 36-month term, about $16,000 over three years, and Western enterprise units later sold in the €17,000–€20,000 band. Any page telling you Pepper "cost $1,800" is quoting about a fifth of the advertised bill.

Can you actually get one in 2026?

The used-buyer checklist matters more for Pepper than for most robots, because parts of the machine depended on servers that no longer answer. First, identify the generation: the classic Choregraphe/Python generation and the later Android-tablet generation (compared here) have different tooling, and cloud-dependent dialog and fleet features may be unavailable on either after Aldebaran's collapse; verify the exact software generation and which network endpoints it needs before paying. Second, ask about the battery (a 795 Wh pack with a claimed ~12-hour retail runtime, long out of production). Third, budget for fragility: university labs reported recurring overheating and brittle shells even before support collapsed. A used Pepper in 2026 is an offline animatronic development platform, not the cloud concierge from the demos.

What Pepper actually did for a decade

Pepper's deployment record is the best argument that it was a phenomenon, not a failure. It greeted customers at HSBC's Fifth Avenue flagship from 2018, staffed Hamazushi restaurants and Carrefour stores, worked Montréal and Prague airports, chanted sutras as a Buddhist funeral priest priced at a fraction of a human one, cheered in empty baseball stadiums during COVID, and taught coding lessons in about a thousand Japanese schools. The one genuinely positive controlled trial is worth knowing: the CARESSES study found that culturally aware Peppers measurably improved older adults' mental health over two weeks of use, with the authors stressing "assist, not replace."

And the famous firing? "Fabio," the Pepper removed from Edinburgh's Margiotta grocery after a week in 2018, was a one-week Heriot-Watt University experiment filmed for a BBC documentary, not an employee. The lead researcher said it plainly to Axios: "'Fired' is not what happened... they thought that was a humorous angle." Pepper still books the occasional real job in 2026: an anonymous donor and a library nonprofit put a $40,000 Pepper to work greeting patrons and leading story time at Honolulu's ʻĀina Haina library last December.

Archival specs

Misinformation audit: seven Pepper claims, checked

Who should buy a Pepper in 2026?

Three buyers, honestly. Labs and HRI researchers: used Peppers at $2,100–$6,800 remain the cheapest full-size social-robot platform ever made, with a huge academic literature behind them, provided you can live on NAOqi 2.5 offline tooling. Venues that want the icon: a working Pepper still draws crowds, as Honolulu's library shows, and RobotLAB's $32,000 supported channel is the lower-risk route (confirm unit condition in writing). Japanese businesses: Pepper+ rental is the only current, supported, manufacturer-backed Pepper on Earth. Everyone else shopping for a working service humanoid in 2026 should read our evidence-ranked guide or the cost guide; the field moved to machines like the Unitree G1 and its successors.

What to watch next

Three things will date this page. Whether Maxtronics ever announces Pepper production, which would flip the "wrong robot" verdict above. Whether Pepper+ leaves Japan; SoftBank has said nothing about export. And NAO7's arrival, which will tell us how serious Maxvision is about the Aldebaran legacy it bought for less than the price of a house in Paris. We re-verify this page against the primary sources above; every figure carries its check date.

The bottom line

Pepper is the most successful failure in robotics: a machine that never worked well enough, sold roughly 17,000 units anyway, defined what the public thinks a social robot is, and earned a Guinness record as the first mass-produced humanoid service robot, awarded five years after its own production line stopped. In 2026 it exists as three different things: a cheap, cloud-orphaned used robot; a $32,000 supported collector's channel; and a Japan-only rental reboot. Know which of the three you're being sold, and Pepper is still, a decade on, exactly what it always was: the friendliest cautionary tale in robotics.

Compare on RoboZaps: Pepper and Unitree G1. Full specifications and a quote route for each, alongside every humanoid robot we carry.